As HPE EVP Neil MacDonald noted before, there are multiple reasons why AI proof of concepts bear little fruit.
A failure to address real-world business needs is one of them. Other stumbling blocks, according to MacDonald, include prohibitive costs, data challenges and misaligned goals.
As was the case in the early days of cloud, if businesses are making AI investment without a clear goal and strategy in place, it’s not terribly surprising that their pilots never make it into production.
At the same time, a 2024 report from PwC found that the majority of South African CEOs recognise the benefits that a tool like generative AI can bring to their business, with some 63% fully expecting it to affect how they deliver value.
So how can businesses make sure that they get this right? The first step: ensuring that all relevant stakeholders within the organisation work together to define what their GenAI future looks like and how they’ll get there. Here are three key questions you need to answer before embarking on any AI initiatives.
In the future, every company that hopes to remain competitive will be an AI-savvy company. The question is, what kind?
For most enterprises, it boils down to a choice between two options. You can be an AI-enabled enterprise, or you can be an AI disruptor.
Up until now, a disproportionate amount of public attention has been focused on the disruptors — the organisations that are spending billions to develop their own large language models (LLMs).
These organisations represent a fraction of companies working with GenAI today.
Most organisations do not have the resources or the expertise to be disruptive at an industry or market level.
Most are simply looking to take existing LLMs and augment them with their own unique enterprise data.
A few more ambitious organisations will be looking to use GenAI to disrupt their internal operating models or re-engineer how they deliver goods and services.
The more data-driven the organisation, the greater the impact this technology will have on how it operates.
At this stage of the game, a lot of corporate interest in GenAI is driven by fear of missing out.
Companies don’t know exactly what advantages they will gain by deploying GenAI, but they’re afraid their competitors will figure it out before they do. Deciding what kind of AI company you want to be, and the kinds of benefits you hope to gain from this technology, is the first step toward moving from experimentation to execution.
Next, you’ll need to tackle the following questions: What are the most pressing problems your business is facing today? Where can GenAI have the greatest impact in solving them?
Only after determining what use cases you want to address with AI should you move on to technical questions:
Despite high adoption rates of AI, successfully completing AI projects remains a challenge for South African businesses.
Companies have a wealth of data, but they’ve struggled to unlock its full potential. It's about turning that mountain of data into actionable insights that drive real change.
That brings us to the third big challenge: How do you pick the right partners to fast-track your GenAI initiatives?
The easy answer is to look for those with an AI-first mentality.
Does the provider offer AI-embedded outcomes in their offerings?
Do these help your business operate in a leaner and more cost-effective way?
Working with vendors that offer AI tools in a standardised and pre-integrated way will allow IT teams to bring GenAI workflows into production more quickly, without having to build everything themselves or figure out which tools work best together.
Especially for organisations with no AI experience, finding the right partners will be a key step, be it to pick the right model, implement quality data pipelines, or deploy applications on the right platforms.
As enterprises experiment with GenAI, they’re gaining valuable experience about what works and what doesn’t, as well as where they need better data or improved governance.
They’re also figuring out which use cases can truly drive the company forward and how to minimise risk while maximising value.
However, to be able to enjoy the enormous productivity gains that come with using these technologies, they need to prioritise strategic and well-planned deployments.
